Ops CLI Documentation
Welcome to the Wasteology Operations CLI (ops) — the unified command-line interface for managing pipelines, databases, projects, reporting, and workflow orchestration across the entire Wasteology portfolio.
Data Pipeline Overview
The Ops CLI monitors and manages the data import pipeline end-to-end:
Quick Start
Installation
The Ops CLI is part of the wg-orchestration project:
Verify Installation
Install tab-completion with ops --install-completion bash (or zsh/fish).
After that, use ops directly — no uv run prefix needed.
Command Reference
All 23 command groups at a glance:
| Command | Description | Common Usage |
|---|---|---|
ops status | Overall system health across all projects | ops status --verbose |
ops pipeline | Monitor and manage data pipelines | ops pipeline list --failed |
ops prefect | Prefect flow and deployment management | ops prefect runs --state FAILED |
ops dbt | dbt data transformation operations | ops dbt run, ops dbt test |
ops sling | SQL Server queries via Sling | ops sling query "SELECT ..." |
ops n8n | n8n workflow automation | ops n8n workflows |
ops digest | Daily digests, standup, task tracking | ops digest standup |
ops portfolio | Cross-project portfolio visibility | ops portfolio status |
ops project | Project registry management | ops project list, ops project health |
ops db | PostgreSQL user and role management | ops db users, ops db grants |
ops agent | Multi-agent orchestration (broadcast, chain) | ops agent broadcast --prompt "..." |
ops ai | Claude and Gemini AI providers | ops ai auto "Summarize this" |
ops orch | ADW orchestrator agent interface | ops orch status, ops orch trigger |
ops monitor | Continuous project monitoring | ops monitor all |
ops docs | Documentation site management | ops docs status, ops docs deploy |
ops tech-stack | Tech stack visualizations | ops tech-stack generate |
ops errors | Recent errors across all projects | ops errors --hours 48 |
ops exec | Execute commands in project context | ops exec wdp-import status |
ops run | Run manifest-registered project commands | ops run dbt run --select staging.* |
ops route | Route tickets to the right project | ops route ticket "Fix billing ETL" |
ops classify | Classify ticket type (bug/feature/chore) | ops classify "Sling timeout on large tables" |
ops query | Query project state for AI assistants | ops query status --json |
ops game | Trash Truck arcade game | ops game |
Common Workflows
Morning Standup
Run ops digest standup to get an AI-generated summary of yesterday's commits across all projects.
Pipeline Triage
Run ops status for a quick overview, then ops prefect runs --state FAILED to see what broke.
Agent Broadcast
Use ops agent broadcast --prompt "..." to fan a question or task out to all project master agents simultaneously.
DB Access
Use ops db create-user + ops db grants to provision access. ops db audit-roles flags missing grants.
The Ops CLI is designed to work with AI agents. Use ops capabilities --json to get a machine-readable dump of all available commands for tool discovery. ops query status --json returns a live system snapshot perfect for piping into an AI context.
Run ops <command> --help in the terminal for full flag documentation on any command, or ops capabilities for a summary of all available commands.
Interactive Architecture Overview
See the interactive architecture overview for a visual map of how the CLI commands connect to data sources, orchestration layers, and downstream systems.
Prefect Single-Surface Lifecycle
The ops prefect group now covers the complete flow lifecycle end-to-end: probe (validate the upstream API) → run-local (smoke-test the flow) → ship / deploy (build image and register the deployment) → run --watch --verify-count (trigger with row-count gating) → doctor (diagnose stalls). No separate Prefect UI or az CLI session required.